Eventos Anais de eventos
COBEM 2023
27th International Congress of Mechanical Engineering
Particle Swarm Optimization and Tikhonov Regularization for source identification over complex regions
Submission Author:
Roseane Albani , RJ
Co-Authors:
Roseane Albani, Juan Pablo de Lima Costa Salazar, Vinicius Albani, Davidson Moreira, Antônio Silva Neto
Presenter: Juan Pablo de Lima Costa Salazar
doi://10.26678/ABCM.COBEM2023.COB2023-0893
Abstract
In this work we propose a source identification modelling that consists of the minimization of the Tikhonov regularization functional using the Particle Swarm Optimization technique. The objective function in the Tikhonov functional considers the discrepancy between the observed and numerical pollutant concentrations. The numerical concentrations are obtained as the solution of an adjoint advection-diffusion partial differential equation, jointly with a non-standard RANS k-$\epsilon$ turbulence model to provide the wind flow field. We perform the simulations using the CEDVAL wind tunnel experimental data A1-1 to provide the flow variables. The model implementation is performed within the framework of the CFD open-source code OpenFOAM and the commercial software Ansys Fluent and their results are compared. The proposed source estimation results were satisfactorily accurate.
Keywords
Source Identification, Particle Swarm Optimization, Computational fluid dynamics (CFD), atmospheric dispersion, OpenFOAM, ANSYS Fluent
